[
https://issues.apache.org/jira/browse/KNOX-587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14697228#comment-14697228
]
Sumit Gupta commented on KNOX-587:
----------------------------------
I see. In the HS2 HA case I was taking the ZK ensemble as input configuration,
querying ZK for the live nodes and picking one at random. I have to keep track
of a HS2 nodes that failed etc, which is similar to the existing HA URL
management code.
BTW, to talk to the ensemble I am using the Apache Curator library
(http://curator.apache.org/).
> Implement a provider for Apache ZooKeeper
> ------------------------------------------
>
> Key: KNOX-587
> URL: https://issues.apache.org/jira/browse/KNOX-587
> Project: Apache Knox
> Issue Type: Sub-task
> Reporter: Kristopher Kane
> Fix For: 0.7.0
>
>
> Implementing Apache Solr will require knowledge of Apache ZooKeeper to
> retrieve the list of live nodes from a Solr Cloud deployment. This sub task
> is to track requirements for a new ZooKeeper provider.
> This provider should not be specific to Solr since other Hadoop ecosystem
> components, that provide a REST API, also use ZooKeeper (HiveServer2). The
> provider should be generic enough to support multiple uses.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
